Introduction Alcoholic hepatitis (AH) is considered a distinct clinical entity in the spectrum of alcoholic liver disease (ALD) with a high short-term mortality. Hence, a number of scoring systems have been developed to assess severity of AH and identify the subgroup of subjects with particularly poor outcome, and suitable for specific treatments or clinical trials. Ability of these models to predict long-term survival is unknown.

Methods Aims: To assess the utility of different published scoring systems in assessing short-term (30 and 90 days) and long-term (6 and 12 months) mortality in patients with AH. Methods We identified patients with liver biopsy evidence of AH from the histology database: 2004–2007. Clinical and histological data were collected from all the electronic records system and from patient notes. The Child Pugh (CP) score, Maddrey discriminant function (mDF), Glasgow AH score (GAHS), Lille model, age, bilirubin, INR, creatinine (ABIC) score, Model of End Stage Liver Disease (MELD) (including its modifications), and UK End-Stage Liver Disease (UKELD) scores were calculated. We evaluated at 30 days, 90 days, 6 months and 1-year outcome. Prognostic utility of the various scores were determined by generating a receiver operating characteristic curve (ROC curve).

Results Of 140 subjects who had transjugular liver biopsy in the study period, 44 had histological evidence of AH. In this subgroup cumulative 30-day mortality was 11.4% (5/44), 90-day mortality was 13.9% (6/43), 6-month mortality was 21.9% (9/41) and 1-year mortality was 26.8% (11/41). Abstract 058 gives the area under ROC curve (AUROC) and 95% CI for the various scoring systems for predicting short and long-term mortality. Abstinence from alcohol (in the first 3 months after diagnosis of AH) was significantly associated with survival at 1 year (p<0.05) and predicted survival with an AUROC of 0.83 (95% CI 0.71 to 0.95)

Conclusion Most scoring systems with the exception of the C-P score appear to be similar in predicting the short-term (30 and 90 days) survival in patients with AH, but were uniformly poor at predicting survival beyond 6 months. Abstinence from drinking within the first 3 months is a better predictor of survival at 1 year.